A5101688594- 3D Shape Modeling and Analysis
- Advanced Vision and Imaging
- Computer Graphics and Visualization Techniques
- Speech and Audio Processing
- Advanced Neural Network Applications
- Music and Audio Processing
- Radiomics and Machine Learning in Medical Imaging
- Quantum Information and Cryptography
- Adversarial Robustness in Machine Learning
- Imbalanced Data Classification Techniques
- Interactive and Immersive Displays
- Advanced MRI Techniques and Applications
- Mobile Learning in Education
- Medical Image Segmentation Techniques
- Domain Adaptation and Few-Shot Learning
- Medical Imaging Techniques and Applications
- Biopolymer Synthesis and Applications
- Lung Cancer Diagnosis and Treatment
- Quantum many-body systems
- Image Retrieval and Classification Techniques
- Computational Drug Discovery Methods
- Impact of Technology on Adolescents
- Photovoltaic System Optimization Techniques
- Human Motion and Animation
- Quantum Computing Algorithms and Architecture
Shanghai Jiao Tong University
2020-2023
Fujian Institute of Oceanography
2023
Ministry of Natural Resources
2023
Massachusetts Institute of Technology
2022
King's College London
2020
Shenyang Normal University
2017
Soochow University
2011
Xiamen University of Technology
2007
There have been considerable debates over 2D and 3D representation learning on medical images. approaches could benefit from large-scale pretraining, whereas they are generally weak in capturing large contexts. natively strong contexts, however few publicly available dataset is diverse enough for universal pretraining. Even hybrid (2D + 3D) approaches, the intrinsic disadvantages within / parts still exist. In this study, we bridge gap between convolutions by reinventing convolutions. We...
This paper addresses the challenging black-box adversarial attack problem, where only classification confidence of a victim model is available. Inspired by consistency visual saliency between different vision models, surrogate expected to improve performance via transferability. By combining transferability-based and query-based attack, we propose surprisingly simple baseline approach (named SimBA++) using model, which significantly outperforms several state-of-the-art methods. Moreover,...
Recent years we have witnessed rapid development in NeRF-based image rendering due to its high quality. However, point clouds is somehow less explored. Compared which suffers from dense spatial sampling, naturally computation intensive, enables deployment mobile computing device. In this work, focus on boosting the quality of with a compact model design. We first analyze adaption volume formulation clouds. Based analysis, simplify NeRF representation mapping function only requires single...
Engineering desired Hamiltonian in quantum many-body systems is essential for applications such as simulation, computation and sensing. Conventional engineering sequences are designed using human intuition based on perturbation theory, which may not describe the optimal solution unable to accommodate complex experimental imperfections. Here we numerically search deep reinforcement learning (DRL) techniques experimentally demonstrate that they outperform celebrated a solid-state nuclear...
We develop an effective point cloud rendering pipeline for novel view synthesis, which enables high fidelity local detail reconstruction, real-time and user-friendly editing. In the heart of our is adaptive frequency modulation module called Adaptive Frequency Net (AFNet), utilizes a hypernetwork to learn texture encoding that consecutively injected into activation layers modulate implicit radiance signal. This mechanism improves expressive ability network with richer basis support, only at...
Diagnosis and treatment of multiple pulmonary nodules are clinically important but challenging. Prior studies on nodule characterization use solitary-nodule approaches nodular patients, which ignores the relations between nodules. In this study, we propose a instance learning (MIL) approach empirically prove benefit to learn By treating from same patient as whole, critical relational information voxels is extracted. To our knowledge, it first study Inspired by recent advances in natural...
Shape primitives decomposition has been an important and long-standing task in 3D shape analysis. Prior arts heavily rely on point clouds or voxel data for extraction, which are less practical real-world scenarios. This paper proposes to learn from multi-view images by introducing implicit surface rendering. It is challenging since shapes have a high degree of freedom, violates the simplicity property primitives. In this work, novel regularization term named Implicit Convexity Regularization...
In order to achieve maximum power of solar cell, we focus on the point tracking (MPPT) algorithm forming based fuzzy control. The control rules are adopted using artificial neural network with measured data. Compared inference systems (FISs) ideal FISs, there is only less than 2% error signal output. simulation conclusions show performance MPPT becomes much precise and active help network.
Emergence of artificial intelligence techniques in biomedical applications urges the researchers to pay more attention on uncertainty quantification (UQ) machine-assisted medical decision making. For classification tasks, prior studies UQ are difficult compare with each other, due lack a unified quantitative evaluation metric. Considering that well-performing models ought know when act incorrectly, we design new metric, area under Confidence-Classification Characteristic curves (AUCCC),...
3D shape analysis has been widely explored in the era of deep learning. Numerous models have developed for various data representation formats, e.g., MeshCNN meshes, PointNet point clouds and VoxNet voxels. In this study, we present Representation-Agnostic Shape Fields (RASF), a generalizable computation-efficient embedding module RASF is implemented with learnable grid multiple channels to store local geometry. Based on RASF, embeddings representations (point clouds, meshes voxels) are...
A new certified reference material (CRM) of squalene (GBW(E) 091179) has been developed using a mass balance (MB) approach in conjunction with quantitative nuclear magnetic resonance (qNMR) technique. In compliance ISO guidelines, the candidate CRM investigated for preparation, structure determination, characterization, homogeneity study, stability studies, as well uncertainty evaluation. The value 99.4%, whose expanded is 0.2% (k=2). Under -20 °C condition, stable no less than 12 months,...
Recent works on learning-based frameworks for Lagrangian (i.e., particle-based) fluid simulation, though bypassing iterative pressure projection via efficient convolution operators, are still time-consuming due to excessive amount of particles. To address this challenge, we propose a dynamic multi-scale gridding method reduce the magnitude elements that have be processed, by observing repeated particle motion patterns within certain consistent regions. Specifically, hierarchically generate...
Wearable electronics are in the news these days and on schedule for Technical Symposium at Display Week 2014.It's hard to miss topic of wearable lately.Entire Web sites, such as CrunchWear.com, dedicated phenomenon.At time this issue was going press, Gizmag.com updating its Electronics department about every other day, with latest entry being a piece LG's still-under-development G Watch, one next-generation "smart watches."IDTechEx had just released new report, "Wearable Technology...
We develop an effective point cloud rendering pipeline for novel view synthesis, which enables high fidelity local detail reconstruction, real-time and user-friendly editing. In the heart of our is adaptive frequency modulation module called Adaptive Frequency Net (AFNet), utilizes a hypernetwork to learn texture encoding that consecutively injected into activation layers modulate implicit radiance signal. This mechanism improves expressive ability network with richer basis support, only at...
Certified reference materials (CRMs) are used to guarantee measurements that follow the international system of units and ensure spatially temporally accurate comparable. In this study, an L-(-)-fucose purity CRM was developed based on ISO principle. The study included structural analysis, characterization, homogeneity test, evaluation stability, estimation uncertainties. mass balance (MB) quantitative proton nuclear magnetic resonance (1H qNMR) methods were for characterizing CRM. Using MB...
Aiming at tree electricity generation key technology, an ideal synchronous rectification structure boost converter is focused on training important variables in sequence of (1) output load R=20O, (2) C and L, (3) especially switch-on RN RP, (4) W/L, (5) area MOSFET. The simulation results suggest that there a CRL (conversion ratio limit) effect, the conversion not only effected by duty ratio, but also adversely proportional to R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML"...
An algorithm based on the active shape model is proposed for segmentation of medical image series. The edge soft tissues in usually are not very clearly, so it's difficult to segment them only by computer. But a tissue's deform sequentially We selected some images from series and segmented with help doctors, results was used as training maps create an PCA Algorithm. this remainder sequence mutual information registration contour model. experiments show efficiency method.
WeChat is the most popular instant message tools in China with its 700 million active users. has been constructed as mobile or blended learning context among student group. This study tried to explore self-efficacy of 7th and 8th graders toward Application Extracurricular Learning Math. Based upon Bandura's social theory, activities instrument will be designed developed, order a teaching strategy for junior school students China.
Abstract Background Cognitive impairment is considered a core feature of psychotic disorders. Deficits in cognition, metacognition and social cognition have been reported to be correlated, indeed predictors, functional outcome or level disability. Psychotic patients tend present lower IQ show specific cognitive domains, than controls. Several studies found deficits facial emotion recognition (FER) higher prevalence the jumping conclusions (JTC) reasoning data gathering biases among patients,...